metadata
library_name: lerobot
tags:
- molmoact2
- robotics
- lerobot
- vla
base_model: allenai/MolmoAct2
molmoact2_block_stack
Fine-tuned MolmoAct2 (action-expert-only) on SO101 data.
| Base model | allenai/MolmoAct2 |
| Dataset | villekuosmanen/armnetbench_block_stack |
| Task | block_stack |
| Training | Vast.ai H100 SXM, batch 64, bf16, gradient checkpointing |
| W&B project | molmoact2_block_stack |
| W&B run | 3uyvgbui |
Checkpoints
| Step | Path |
|---|---|
| 006000 | checkpoints/006000/pretrained_model/ |
| 004000 | checkpoints/004000/pretrained_model/ |
Usage
from lerobot.policies.molmoact2.modeling_molmoact2 import MolmoAct2Policy
policy = MolmoAct2Policy.from_pretrained("pravsels/molmoact2_block_stack", revision="main")